H.I. No. 8: First World YouTuber Problems: Difference between revisions
H.I. No. 8: First World YouTuber Problems (edit)
Revision as of 03:00, 15 December 2022
, 1 year ago→Other: Cleaned transcript up to 20:30
Timscriber (talk | contribs) (→Other: Cleaned transcript up to 15:36) |
Timscriber (talk | contribs) (→Other: Cleaned transcript up to 20:30) |
||
Line 374:
Brady: --I haven't thought this through. [chuckles] It, uh, if anyone decides that they're going to do that, I don't know, I mean, maybe I've- no, don't email me.
Grey: You know, you're still gonna read 'em. [chuckles]
Brady: Anyway, I did get this email from a chap named Sean... I don't know if I should say his surname, 'cause he might like it being read out, but I'm always really reluctant to use people surnames, um, publicly.
Grey: Uh, yeah, I'd- I'd say the general policy let's stick with- with first names, right? That's what I would--
Brady: Sean, you know who you are, you wrote a long email. It was very good, you made a lot of good points. Some I agree with, some I have no opinion on--
Grey: Mm-hmm.
Brady: --uh, but some I didn't make in our discussion.
Grey
Brady: I mean, I- like most things in life, I have no really firm position on this whole language-in-school debate. But, I tried to put the other side because... you put- you put your side so strongly, and he made some really good points, and I'd like to share a couple of them with you. One of the points he made, you talked about in the last podcast, how you think, sort of, machine translation is moving along so quickly,--
Grey: Mm-hmm?
Brady: --that that's one of the reasons speaking other languages will become sort of, you know, a less important skill--
Grey
Brady: --as time goes by.
Grey: Mm-hmm.
Brady: He's- he politely but strongly disagreed with that. And he seems- it seems to be an area he knows quite a lot about.
Grey: Mm-hmm.
Brady: And he- he thinks were fifty, if not a hundred years away from computerized language translation being really top notch, like reaching a level that, um, is acceptable for, you know, strong purposes.
Grey: Mm-hmm.
Brady: I'll send- I'll send you the email,- and you can decide if you want to read it.
Grey: Mm-hmm.
Brady: He made a strong case for it.
Grey: Yeah.
Brady: So I just want to throw that in the mix. I don't- I don't know if he's right or not,--
Grey: I- I- my- my- my, uh, my thought on that just- just very briefly is, when I talked about the- the language translation, we- we discussed it a little bit last time, but my main thought about that is the scenario of- of utility, right? It is a good enough for most people traveling in a foreign country to communicate with the people around them. And, I think that that is the language scenario for most people. I- I- I imagine, especially, especially if you are ever having to, uh, translate a- a work that has any kind of nuance to it, like if you're translating a novel between languages, I don't- I don't think machine translation is going to be able to do that kind of work for a very long time, because I think that's a- that- that- that is not a question of- of communication, uh, translating something that has style to it. [clears throat] Excuse me. Um, I think--
Brady: We'll blame Derek for that cough.
Grey: Yeah, yeah, I will blame Derek for that cough. What I was just gonna say is- is, uh, I was reading a document a while ago, talking about, uh, the beginning points of- of machine language translation, like, what- "how do you begin to train these algorithms?", and I think to give you a sense of how- of how dry and utilitarian the translations may be, is that the- the starting point for this one algorithm that I was reading about was the European Union Corpus of Laws. Because, uh, for- for those of- of you who are aware, in the European Union (oh god, I should know this off the top my head), but I think it's there's something like twenty-two or twenty-four different official languages,--
Brady: Hmm.
Grey: --and what that means is that any document that is issued in the European Union has to have legal copies in all of the other languages.
Brady: Nice.
Grey: Which is, you can find these- these statistics for how many human hours is employed in--
Brady: [chuckles]
Grey: --uh, in this endeavor, which is- which is quite a lot. For the record, I just want to say I'm totally okay with that, right? If you're going to designate something as an official language, you need to produce documents in that language, otherwise, what's the point?
Brady: Yeah.
Grey: But basically, the result is that you end up having just this enormous treasure trove for computer linguistics of data of thousands of documents that have been translated into this twenty-something different languages. And so it's- it's a great training ground. And by- by training I mean, algorithms a little weird, you can actually sort of have them learn on their own. It's- but, the- the result means that, like, your translation style is going to sound probably as dry as a European Union legal memo is going to sound. Uh, so I don't- I don't think that we're going to have translations of poetry that are anything but laughable for quite a long time.
Brady: Well, if that's your- if that's your brief response to that first of Sean's points,--
Grey: Oh god... [laughing]
Brady: --I'll, uh, I'll- [laughing] I'll limit the number of Sean's points I share with you.
Grey: Oh, I'm so sorry.
Brady: And I will point out, I'm only briefly summarizing, I'm not- I'm not pointing out his whole argument.
Grey: Yes, yes.
Brady: Another point he made it was very good though, um, you were saying you're an advocate of coding, computer programming--
Grey: Mm-hmm.
Brady: --as something that would be more valuable to teach in schools.
Grey: Mm-hmm.
Brady: He rightly points out, computer languages are- change very quickly become outdated very quickly.
Grey: Mm-hmm.
Brady: Uh, you know, what- what languages are we going to- what computer languages are we going to teach in school? Are people going to come out fluent in some code that is redundant before they even reach the job market?
Grey: Yeah, so this is- I- I think- it's an interesting point, I actually, uh, discussed this with Derek a- a little bit, uh, tonight. In any future video or any- any, uh, future podcast that we talk about, I mean computer programming languages in a very very broad way. I- I think there are- there are lots of things if I- if I was having a- a computer language course, that I- that I would teach that would not be straight up computer languages.
[20:30]
So, um, for example, when I- when I worked as a teacher, I taught a course in electronics, and that was a 100% hands on practical course no code at all. But what it was was the was kids had to Connect a bunch of sensors and buttons and switches to each other. And they had to make those connections using logic gates. So those are things like, if the heat sensor is active, and the light sensor is active, then make the buzzer sound. So so I think any kind of, of teaching of logic to me falls under this category of what a computer programming languages, just the very idea of how do you talk to computers. And so even though computer programmers in the audience, right, their head will explode when I put this in the same category. But I would be totally happy to have a lot done in say, spreadsheets. There's an enormous amount of stuff you can do with a spreadsheet that is if then else, right, if this number is this, then do that thing. That's what I mean by computer programming, just the idea of teaching, how to communicate with computer In a logical way, I just want to say that that's that's a very, very broad, very broad definition. But to specifically answer his question about the the language, I would just say that the language is in the field seem to turn over very slowly. There are languages that don't move very much. And okay. For example, if you're using and it's, it's 2014. And if you're using an iPhone, those iPhones are still programmed in, basically in in a variant of the language of C. Which God I mean, off the top of my head, I'm going to say C is at least 1970s, kind of old. So that's not I mean, sees not a pretty language I've programmed a little bit and see, sorry, see programmers, it just kind of ugly. Lisp is beautiful for those in the audience who are listening, and I do like Python. But anyway, you can pick languages that don't change very much but much more with computer languages than with say, human languages, the the internal concepts, the logical gates of if then else. And NAND, those things are universal and languages and the syntactic differences between programming languages are not as large as you might think they are. So
Brady 23:16
| |||